aboutsummaryrefslogtreecommitdiff
path: root/CHANGES
Commit message (Collapse)AuthorAge
...
* Changes for 0.9.Nicholas Marriott2009-07-01
|
* This file is not going to be used as much any longer, make a note.Nicholas Marriott2009-07-01
|
* stat(2) files before trying to load them to avoid problems, for example with ↵Nicholas Marriott2009-05-21
| | | | | | | "source-file /dev/zero". This commit dedicated to Tom: protecting idiots from their own stupidity for more than 20 years.
* Try to guess if the window is UTF-8 by outputting a three-byte UTF-8 wide ↵Nicholas Marriott2009-05-19
| | | | character and seeing how much the cursor moves. Currently tries to figure out if this works by some stupid checks on the terminal, these need to be rethought. Also might be better using a width 1 character rather than width 2.
* If LANG contains "UTF-8", assume the terminal supports UTF-8, on the grounds ↵Nicholas Marriott2009-05-19
| | | | that anyone who configures it probably wants UTF-8. Not certain if this is a perfect idea but let's see if it causes any problems.
* - New window option monitor-content to search for a string in a window, andTiago Cunha2009-05-19
| | | | | | | | | highlight the status line if it matches. - To make this possible, the function cmd_find_window_search from cmd-find-window.c had to be moved to window.c and renamed window_pane_search. - While there use three new functions in server.c to check for bell, activity, and content, to avoid too much nesting.
* main-horizontal layout and main-pane-height option to match vertical.Nicholas Marriott2009-05-18
|
* left-vertical -> main-vertical.Nicholas Marriott2009-05-18
| | | | | Also update CHANGES/TODO.
* select-layout command and some key bindings.Nicholas Marriott2009-05-16
|
* Recreate server socket on SIGUSR1, per SF feature request 2792533.Nicholas Marriott2009-05-16
|
* Keys in status line (p in vi mode, M-y in emacs) to paste the first line of ↵Nicholas Marriott2009-05-14
| | | | the upper paste buffer. Suggested by Dan Colish.
* clear-history command.Nicholas Marriott2009-05-14
|
* Don't force wrapping with \n when asked, let the cursor code figure it out. ↵Nicholas Marriott2009-05-14
| | | | Should fix terminals which use this to detect line breaks.
* Add a changes entry.Nicholas Marriott2009-05-13
|
* Space trimmage mega-diff.Nicholas Marriott2009-05-04
|
* Use ACS for line drawing characters.Nicholas Marriott2009-05-04
|
* Support "neww; neww" as well "neww ; neww".Nicholas Marriott2009-04-30
|
* previous-layout command.Nicholas Marriott2009-04-30
|
* Use a (pre)randomised binary tree for UTF-8 character widths. Probably overkillNicholas Marriott2009-04-30
| | | | | but meh.
* Merge resize-pane-{up,down} into resize-pane.Nicholas Marriott2009-04-30
|
* Some tweaks for Solaris.Nicholas Marriott2009-04-29
| | | | | Get rid of vis.* in favour of a small replacement function.
* Avoid using the prompt history when the server is locked, and prevent anyTiago Cunha2009-04-27
| | | | | input entered from being added to the client's prompt history. From nicm.
* - confirm-before command.Tiago Cunha2009-04-27
| | | | | | - Bound "&" and "x" by default to confirm-before "kill-window" and confirm-before "kill-pane", respectively.
* Support NEL.Nicholas Marriott2009-04-23
|
* Use the xenl terminfo flag to detect early-wrap terminals like the FreeBSDNicholas Marriott2009-04-23
| | | | | console. Many thanks for a very informative email from Christian Weisgerber.
* Note it.Nicholas Marriott2009-04-23
|
* Remove the right number of characters from the buffer when escape then a cursorNicholas Marriott2009-04-17
| | | | | key (or other key prefixed by \033) is pressed. Reported by Stuart Henderson.
* rotate-window command.Nicholas Marriott2009-04-03
|
* Change scroll/pane redraws to only redraw the single pane affected rather thanNicholas Marriott2009-04-02
| | | | | the entire window.
* If redrawing the region would mean redrawing > half the pane, just schedule toNicholas Marriott2009-04-02
| | | | | | | redraw the entire window. Also add a flag to skip updating the window any further if it is scheduled to be redrawn. This has the effect of batching multiple redraws together.
* Tweak to CHANGES.Nicholas Marriott2009-04-01
|
* Basic horizontal splitting and layout management. Still some redraw and otherNicholas Marriott2009-04-01
| | | | | | | | | | | issues - particularly, don't mix with manual pane resizing and be careful when viewing from multiple clients; generally cycling the layout a few times will fix most problems. Getting this in for testing while I think about how to deal with manual mode. Split window as normal and cycle the layouts with C-b space. Some of the layouts will work better when swap-pane comes along.
* +UTF-8 fixed to CHANGES.Nicholas Marriott2009-03-31
|
* Update CHANGES.Nicholas Marriott2009-03-31
|
* Draw vertical line on the right when window size is limited.Nicholas Marriott2009-03-31
|
* Better UTF-8 support, including combined characters. Unicode data is now storedNicholas Marriott2009-03-28
| | | | | | | | | | | as UTF-8 in a separate array, the code does a lookup into this every time it gets to a UTF-8 cell. Zero width characters are just appended onto the UTF-8 data for the previous cell. This also means that almost no bytes extra are wasted non-Unicode data (yay). Still some oddities, such as copy mode skips over wide characters in a strange way, and the code could do with some tidying.
* Key repeating is now a property of the key binding not of the command. RepeatNicholas Marriott2009-03-28
| | | | | | is turned on when the key is bound with the -r flag to bind-key. next/previous- window no longer repeat by default as it turned out to annoy me.
* Clear using ED when redrawing the screen. I foolishly assumed using spacesNicholas Marriott2009-03-28
| | | | | | would be equivalent and terminals would pick up on this, but apparently not. This fixes copy and paste in xterm/rxvt.
* Sockets in /tmp are now created in a subdirectory named, tmux-UID, egNicholas Marriott2009-03-27
| | | | | | | | | | tmux-1000. The default socket is thus /tmp/tmux-UID/default. To start a separate server, the new -L command line option should be used: this creates a socket in the same directory with a different name ("-L main" will create socket called "main"). -S should only be used to place the socket outside /tmp. This makes sockets a little more secure and a bit more convenient to use multiple servers.
* * New session flag "set-remain-on-exit" to set remain-on-exit flag for new ↵Nicholas Marriott2009-03-21
| | | | windows created in that session (like "remain-by-default" used to do). Not perfectly happy about this, but until I can think of a good way to introduce it generically (maybe a set of options in the session) this will do. Fixes SF request 2527847.
* Support for 88 colour terminals.Nicholas Marriott2009-03-07
|
* break-pane command to split a pane off into a new window; bound to ! by default.Nicholas Marriott2009-03-07
|
* Sort out escape key handling so it works.Nicholas Marriott2009-03-02
|
* Redo mode keys slightly more cleanly and apply them to command prompt ↵Nicholas Marriott2009-02-13
| | | | editing. vi or emacs mode is controlled by the session option status-keys.
* Looking up argv[0] is expensive, so just use p_comm for the window name ↵Nicholas Marriott2009-02-13
| | | | which is good enough. Also increase name update time to 500 ms.
* Only use RI at the top of the screen.Nicholas Marriott2009-02-12
|
* FreeBSD's console wraps lines at $COLUMNS - 1 rather than $COLUMNS (the ↵Nicholas Marriott2009-02-11
| | | | cursor can never be beyond $COLUMNS - 1) and does not appear to support changing this behaviour, or any of the obvious possibilities (turning off right margin wrapping, insert mode). This is irritating, most notably because it impossible to write to the very bottom-right of the screen without scrolling. To work around this, if built on FreeBSD and run with a "cons" $TERM, the bottom-right cell on the screen is omitted.
* Emulate CSR with existing screen data on old/stupid terminals which don't ↵Nicholas Marriott2009-02-11
| | | | support it.
* Don't redraw status line unless it has actually changed. Stops extraneousNicholas Marriott2009-02-10
| | | | | | updates between clock/#() changes and doesn't require manual status-interval 0 when no updates are occuring.
* Handle "" properly.Nicholas Marriott2009-02-08
|